Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Rambus (RMBS) - Eagle or Penguin -- Ignore unavailable to you. Want to Upgrade?

To: Bilow who wrote (40216)	4/18/2000 7:41:00 AM
From: John Walliker	Respond to of 93625

Carl, It isn't a matter of the user making any calculations, his operating manual tells him what to do. If he inserts the wrong parts, then the BIOS tells him to yank it out. If doing this is acceptable for DDR memory, I don't see why it should not have been acceptable for Rambus. Rambus comes in a bunch of different flavors But not very many. The only differences are in speed bins and whether ECC is available. More combinations are possible when the devices are combined onto a RIMM, but the same applies to any other memory. Re the reason for requiring registered DIMMs. This doesn't matter. Just follow the rules given by the manufacturer. Don't put EDO DRAM in a SDRAM only system, for instance. No big deal. The amazing thing is that they could make a motherboard that is compatible with both registered and unregistered DRAM. It may not matter to the computer buyer, but it does matter to an informed discussion of signal integrity issues. I would suggest that ringing becomes intolerable when multiple non-registered DIMMs are used. Can you offer evidence to contradict that assertion? Why should it be surprising that they can make a motherboard that supports both types. After all, it is not possible to terminate a network with multiple stubs accurately when the number of devices inserted is unknown. Therefore, anything which reduces the stub length will improve matters. Maybe the eventual marketing solution, since most machines are sold with some memory installed, is to only install 2 DIMM slots in MBs intended to be unregistered, and put 3 in the registered machines. What's the big deal? Registered DIMMs are going to cost more. I'll link in a quote in a minute, if I can quickly find one. So why was it such a big deal when Intel restricted the number of slots that could be used with Rambus? You seem to be suggesting that some of the restrictions that have been associated with Rambus might also apply to DDR. That was not your earlier position. By the way, what say you about my recent post on Intel's i840 RDRAM power consumption notes, as compared with the hype at dramreview? ( #reply-13441932 ) It is well known that faster memory types will dissipate more power when active. It is also well known that Rambus can concentrate the heat dissipation in one chip under some conditions. I showed a long time ago that when comparing well terminated systems the Rambus system power was lower than the alternatives. John